SEARCH

SEARCH BY CITATION

Keywords:

  • conservation;
  • corridor;
  • fragmentation;
  • metapopulation;
  • remnants

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

1. In order to determine the influence of habitat shape on aspects of the population dynamics of birds, 196 male red-capped robins Petroica goodenovii were surveyed using song playback in large non-linear woodland remnants and linear roadside remnants in the wheatbelt of south-eastern Australia.

2. The average density of male robins was significantly higher in large non-linear remnants [0·86 ± 0·09 (SD) birds ha–1] than in small linear remnants [0·35 ± 0·16 (SD) birds ha–1].

3. Red-capped robins exhibit delayed plumage maturation, and 14% of the males captured were yearlings. This percentage was significantly higher in roadside remnants (20%) than in large, non-linear remnants (8%).

4. These results indicate that different population processes are occurring in the individual remnants and the dynamics of the metapopulation are potentially complex.

5. Land managers must not focus excessively on wildlife corridors (narrow, linear habitats) at the expense of appropriate management and restoration of large areas of native vegetation.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Habitat loss is the most important cause of decline of global biodiversity (UNEP 1995). Consequently, in many landscapes in which the dominant land use is agricultural production, reconstruction of native vegetation is now recognized as a priority, both for conservation of biodiversity and for maintaining landscape productivity (Saunders, Hobbs & Ehrlich 1993). The configuration (area, shape and position) of reconstructed habitat patches is an important consideration in their design. At present, guidelines for configurations for soil and catchment protection are more readily available to land managers than guidelines for reconstruction of wildlife habitat. Furthermore, guidelines for construction of wildlife habitat continue to be dominated by wildlife corridors, despite considerable debate in scientific circles about their value (Hobbs 1992; Simberloff et al. 1992). There remains a need for research into the relative value of different configurations of vegetation for wildlife habitat as well as for conduits of movement.

Birds have been the focus of much of the research on the effects of habitat loss on fauna. Surveys of birds in fragmented rural landscapes in several regions of Australia have indicated that there is a suite of species that continue to decline (Saunders 1989; Robinson 1991; Barrett, Ford & Recher 1994) even after the initial loss of habitat has occurred. Ground foraging and nesting species appear to be particularly vulnerable (Recher & Lim 1990; Robinson 1991; Garnett 1992). There are two main, non-exclusive, explanations for this continuing decline.

The first is a function of habitat area and derives from the fact that small patches of habitat support only small subpopulations of birds. Small subpopulations are vulnerable to local extinction from stochastic factors, such as drought or overgrazing (Robinson 1993). If the probability of recolonization from neighbouring habitat patches is reduced through isolation, local extinction of a subpopulation may be permanent. Regional extinctions may follow from a series of local extinctions (Robinson 1993). Ground nesting or foraging species may be expected to be at particular risk because stochastic factors associated with agriculture and pastoralism have their greatest impact at ground level.

The second is a function of edge-to-area ratio, because fragmentation of habitat results in an increase in edge habitat. Whereas much of rural Australia was once a continuous expanse of woodland in which grassland–woodland ecotones were relatively uncommon, much of the remnant vegetation in regions such as the New South Wales and Western Australian wheatbelts is now ecotonal. Most remnants are small or exist as linear strips along roadsides (Lynch & Saunders 1991; Sivertsen & Metcalfe 1995) and both these configurations have high edge-to-area ratios (Sisk & Margules 1993). Edge and interior habitats differ in physical and ecological characteristics, including wind velocity, light level, temperature, weed invasion and the presence of flora and fauna that are edge specialists (Murcia 1995). In particular, numerous studies have shown that the impact of some predators (Gates & Gysel 1978; Møller 1989; Gibbs 1991) is higher along edges. The resultant increase in adult and/or nest mortality with increasing edge and decreasing area should lead to higher turnover of individuals. A higher turnover has two consequences.

First, if supply of replacement individuals is limited, then some patches of suitable habitat will remain vacant when gaps occur as a result of death of territory holders. The supply of replacement individuals will be limited when a patch is isolated and/or reproductive output is depressed at a local and/or landscape scale. Vacancies will be reflected in a lower population density in sites that are more isolated or in which reproductive output is lower.

Secondly, there are consequences for the age structure of subpopulations because replacement birds are more likely to be yearling birds. The dominant dispersive phase for most passerine birds is during the first year of life (Greenwood & Harvey 1982; Donovan et al. 1995a) so that replacements resulting from immigration are more likely to be young. Replacements resulting from on-site reproduction will necessarily be young. The age structure should thus be lower in small or linear habitat patches.

The aim of this study was to determine whether these predicted consequences of habitat fragmentation hold for a small passerine. In particular, we compared the age structure and density of populations found in large continuous patches of habitat with those occupying linear strips of vegetation along roadsides.

Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Study species

The red-capped robin Petroica goodenovii (Vigors and Horsfield) is a member of the Australo-Papuan Robin family Eopsaltridae and is a typical member of the Petroica genus. Red-capped robins weigh approximately 9 g and are largely insectivorous ground feeders (Frith 1983). Their plumage shows distinct sexual dimorphism and males exhibit delayed plumage maturation (Boles 1988). Males in adult plumage are black above, with a bright scarlet forehead and breast. In contrast, females and yearling males are indistinguishable, with upper parts and breast of buff-grey, a reddish forehead and varying amounts of a reddish wash on the breast.

Red-capped robins occur in lightly timbered areas such as open woodland, mallee and mulga scrubs (Boles 1988). They are sedentary, although they may be partially nomadic, with birds in winter occupying roadside vegetation in which they are absent during summer (Lynch & Saunders 1991). Although generally common and widespread throughout their range, declines in red-capped robin populations have been documented in various parts of Australia (Hoskin, Hindwood & McGill 1991; Robinson 1993; Saunders & Ingram 1995; Egan, Farrell & Pepper-Edwards 1997).

Both adult and yearling males defend breeding territories, attract mates and breed (Boles 1988). They both respond strongly to playback of territorial song by approaching the source and singing a territorial song and emitting a ‘ticking’ call. Female robins sometimes respond to taped territorial song by approaching and ‘ticking’, but they do not sing.

Study sites

This study was carried out in 10 woodland remnants near Forbes in the wheatbelt of New South Wales (Fig. 1). Five linear remnants were located in travelling stock routes or roadside reserves between 50 and 90 m wide and between 40 and 105 ha in area. The other five remnants were large, non-linear state forests of between 640 and 1860 ha in area. Hereafter we refer to these two types of remnant as linear remnants and large remnants, acknowledging that there is an interrelationship between remnant shape and remnant size.

image

Figure 1. Location of study sites (numbered) in the wheatbelt of New South Wales, Australia. All vegetation remnants shown are Peneplain Box Woodlands or Peneplain White Cypress Pine Woodlands. The inset map shows the study region shaded dark inside a boundary indicating the extent of the New South Wales wheatbelt.

Download figure to PowerPoint

Sites were chosen on the basis that they contained red-capped robins and had vegetation as similar as possible to each other. All selected sites have been classified by Sivertsen & Metcalfe (1995) as either Peneplain Box Woodlands or Peneplain White Cypress Pine Woodlands. All sites had Callitris glaucophylla J. Thompson & L. Johnson as the dominant tree and understorey species. Other common tree species included Eucalyptus populnea F. Muell., E. microcarpa (Maiden) Maiden, E. intertexta (R. T. Bak.), Brachychiton populneus (Schott) R. BR. and Allocasuarina luehmannii R. T. Bak. Common shrubs included Acacia deanei (R. T. Bak.) and Dodonaea viscosa (L.) Jacq.

There was evidence of grazing at most remnants and it is likely that all remnants had been grazed in the recent past.

Bird survey

A 500 × 500-m grid, marked at 50-m intervals, was established in each of the large remnants. Song playback was conducted at alternate points on the grid such that a playback point was situated every 100 m. Playback points were located a minimum of 50 m from the boundary of the grid, resulting in 25 playback points in each large remnant. In linear sites, no grid was established: song playback was conducted from points at 100-m intervals along the mid-line of the strip.

Up to four 15-s bursts of song, interrupted by 15-s listening periods, were played at each point from a hand-held tape recorder. Playback was terminated at each point immediately upon detecting a red-capped robin. If a bird responded to the playback, a mistnet was erected near the playback point. Speakers were set up on either side of the net and song was played again. By alternating the song between one speaker and the other, the target bird could be induced to fly into the net as it transferred its attention between speakers. All birds captured were colour banded with unique combinations.

Males in adult plumage were easily sexed from their plumage. Yearlings were sexed on the basis of their song. In most cases, pairs of birds were attracted by the playback and so the appearance of a pair in ‘female’ plumage was a good indication that one bird was a putative yearling male. Birds in ‘female’ plumage were only sexed as males if they were induced to sing, which sometimes necessitated repeated playback.

Survey and capture of all male red-capped robins within a site took up to 4 days. Following survey and capture of all birds in a site, the site was resurveyed using playback to verify that no unbanded males remained. One linear remnant and one large remnant were surveyed each month between August and December 1995. Red-capped robins nested in all our chosen remnants during the period of the study.

The number of adult and yearling males was determined for each site, and differences in their relative frequencies between large remnants and linear remnants were determined by chi-square analysis. The total number of males in each site was converted to density, and differences in density between large and linear remnants were determined by single factor anova.

Habitat survey

In each site, vegetation was assessed along five 250 × 1-m transects. We recorded the number of individuals of each species of tree (≥ 5 m) and shrub (< 5 m), and estimated percentage ground cover of grass, leaf litter and bare earth in five 1 × 1-m quadrats along each transect. Data from the five quadrats were pooled to provide a single estimate of each type of ground cover for each transect. Data on vegetation and ground cover for each transect were used as replicates, in a series of two-factor nested anovas, to determine whether there were significant differences in each habitat variable between large and linear remnants. Non-metric multidimensional scaling and analysis of similarity (ANOSIM; Clarke & Warwick 1994) were used to determine whether large remnants could be separated from linear remnants using a combination of habitat variables.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Habitat variability

The abundance of all tree and shrub species other than Callitris glaucophylla was too low to be analysed individually. For the purpose of analysis, all Eucalyptus species were pooled into a single category, as were all shrubs less than 5 m tall, other than C. glaucophylla.

There was a high degree of variability in habitat characteristics between different sites (Table 1). The differences between sites were significant for all characteristics except the percentage cover of bare earth, but there were no consistent differences between large and linear remnants (Table 1).

Table 1.  Results of nested analyses of variance testing for differences in seven habitat characteristics between five large and five linear remnants
Large vs. linearBetween remnants
Habitat variableFd.f.PFd.f.P
Callitris < 5 m0·1371,80·7213·908,550·001
Other shrubs < 5 m0·5671,80·5693·548,550·002
Callitris ≥ 5 m0·0071,80·9344·988,550·000
Eucalypts0·0211,80·8878·048,550·000
Grass0·0091,80·9975·508,550·000
Leaf litter0·0391,80·8493·228,550·004
Bare earth0·0331,80·8601·298,550·266

Non-metric multidimensional scaling and analysis of similarity failed to separate large sites from linear sites when all habitat variables were considered in the same analysis (Fig. 2). It is noteworthy (see the Discussion) that Remnant 5 was the most different of all the 10 sites (Fig. 2). Remnant 5 had a low density of all tree and shrub species and a high percentage cover of grass. The cover of leaf litter was correspondingly low.

image

Figure 2. MDS plot showing degree of similarity in habitat characteristics between each of the 10 study remnants. Remnants that are located closer together have greater habitat similarity. Remnants 1–5 are large remnants and 6–10 are linear remnants. No significant difference in habitat characteristics was detected by ANOSIM (Global R = 0·092, P = 0·159).

Download figure to PowerPoint

Bird density

The density of male red-capped robins was extremely consistent between large remnants despite the variation in habitat (Fig. 3). The number of males captured in the five 25-ha plots ranged from 18 to 24 with a mean of 21·6 ± 2·3 (SD) birds. This corresponds to a mean density of male red-capped robins in large remnants of 0·86 ± 0·09 (SD) males ha–1.

image

Figure 3. Variation in the density of male red-capped robins between linear remnants and large remnants.

Download figure to PowerPoint

The area surveyed in linear remnants was not constant; we continued to capture birds within a remnant until 20 males had been captured or until we had sampled the whole remnant. The mean density of male red-capped robins in linear remnants ranged from 0·18 to 0·53 with a mean of 0·35 ± 0·16 (SD) males ha–1. The density in linear remnants was thus significantly lower than in large remnants (F = 38·25, d.f. = 1,9, P < 0·001).

Age structure

Of the 196 male red-capped robins that were captured, 14% were yearlings. There was a significantly higher proportion of yearling males in linear remnants than in large remnants (χ2 = 4·09, d.f. = 1, P = 0·043). Twenty per cent of singing males in linear remnants were yearlings, compared with 8% in large remnants.

Although small samples prevented statistical comparisons of the proportion of yearling males in individual sites, there was considerable variation between sites within each remnant type (Fig. 4). In particular, Remnant 5 had a relatively high proportion of yearling males.

image

Figure 4. Variation in the percentage of all male red-capped robins that were yearlings between linear remnants and large remnants.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Few studies have measured demographic parameters of the same species in different habitats, or in different-sized patches of the same habitat (Pulliam & Danielson 1991; Villard, Martin & Drummond 1993). This study has demonstrated that the density and age structure of populations occupying small linear remnants of habitat were different from those of populations occupying large, non-linear remnants. This indicates that different population processes were occurring in the individual remnants and the dynamics of the metapopulation is potentially complex.

Other research (Temple & Carey 1988; Donovan et al. 1995a,b) has demonstrated that small remnants may act as population sinks, draining source populations. Depending on demographic parameters, population sinks may or may not contribute to the stability of the metapopulation (Howe & Davis 1991; Pulliam & Danielson 1991; Donovan et al. 1995a). The younger age structure and lower density of the populations occupying small linear remnants in this study suggest that these remnants were inferior, but without information on whether recruits to a remnant were sourced from the remnant itself, we cannot determine the degree to which they supply or drain the metapopulation. However, designers of programmes of vegetation reconstruction should be aware that establishing relatively small linear strips of vegetation will not achieve the same result for wildlife conservation as protecting existing remnants.

We are unable to attribute the differences in density and age structure to linearity per se because the shape of remnants (linear vs. non-linear) was necessarily confounded with the size of remnants. For historical reasons, linear and non-linear remnants of the same size and vegetation type do not exist in the wheatbelt of NSW. The result of this confounding is that the banded populations from linear remnants were small, discrete subpopulations and the banded populations from non-linear remnants were subsets of larger subpopulations. Linear remnants were thus more isolated from dispersers than large remnants and the associated differences in immigration and emigration can potentially explain the lower population density.

Remnant size can also mediate an effect on adult survival through small population size. For example, for both probabilistic reasons associated with small populations, as well as deterministic reasons associated with female choice (Villard, Martin & Drummond 1993), males may experience difficulty finding mates. If unpaired birds undergo higher stress because of prolonged territorial advertisement, they may experience greater mortality or increased emigration. This would result in lower population density and a younger age structure, consistent with our observations.

Our results can also be explained by changes to breeding productivity and adult mortality correlated with habitat size and/or shape. If adult survival or breeding productivity is negatively influenced by edge habitat (Andrén & Angelstam 1988; Ratti & Reese 1988; Møller 1989; Gardner 1998), density is likely to be lower in small, linear remnants. However, reduced breeding productivity alone cannot explain our results because the age structure would become older, all else being equal. A younger age structure in association with decreased density can only eventuate from increased adult mortality or emigration, with territory vacancies filled by reproduction or juvenile dispersal. Juvenile dispersal is typically the dominant dispersal phase for passerine birds (Greenwood & Harvey 1982; Donovan et al. 1995a).

An increased proportion of edge habitat could reasonably be hypothesized to result in increased adult mortality in linear remnants. For example, increased predation on adults by predators, which are more effective or abundant along edges, could operate in much the same way as it has been shown to operate for predation on eggs and nestlings (Andrén & Angelstam 1988; Ratti & Reese 1988; Møller 1989; Gardner 1998). Several of the species of predator responsible for increased nest predation in linear remnants in the same study sites (Major et al. 1999) are also known to be predators of adult birds, e.g. pied butcherbird Cracticus nigrogularis (Gould) and grey butcherbird Cracticus torquatus (Latham). Small birds might also be less able to escape predators in linear remnants than in the more extensive woodland provided by large remnants.

Poorer habitat quality could also lead to increased adult mortality or emigration and increased juvenile immigration. As younger birds are generally competitively inferior (Sherry & Holmes 1989), older birds may occupy the available area of high-quality habitat in large remnants, leaving a greater proportion of inferior, linear habitat for yearling birds. Depending on reproductive output of the metapopulation, the inferior habitat may be occupied at a lower density. However, our assessments of vegetation demonstrated that although there was considerable variation between remnants in all except one of the variables we measured, there was no indication of a consistent difference between large and linear remnants. We therefore conclude that vegetation, and consequently ground cover, does not explain differences in density and age structure of red-capped robins between large and linear remnants.

It is worth noting, however, that one large remnant (Remnant 5) stood out from the other large remnants in having a high percentage of yearlings, although still having a comparable density of males. This remnant had the most different vegetation of any of the 10 sites. The study grid for Remnant 5 was in a portion of a state forest dominated by a young stand of Callitris glaucophylla with very little other vegetation and an extremely dense grass understorey. A possible explanation for the difference in age structure of this site is that this particular part of the forest was of low quality and acted as a sink for young males produced in higher-quality parts of the forest.

The presence of roads through the length of the site is another way in which linear remnants differed from large remnants and this could potentially cause the increase in adult mortality required to explain our results. Two roads were major highways but the other three were minor roads that would experience only a few vehicle movements each day. Birds were frequently observed to cross roads, and it is possible that collision with vehicles could result in higher mortality in linear remnants. Interestingly, the two linear remnants with the highest percentage of yearling males (Remnants 8 and 9) were those that were contiguous with the two highways.

In conclusion, this study found that in small linear remnants the density of male red-capped robins was lower, and the age structure younger, than in populations occupying large non-linear remnants. Thus, even though the linear strips we studied were relatively broad in the context of the New South Wales wheatbelt, they supported different population processes to large remnants. It is reasonable to expect that narrower strips, such as those typically planted along roadsides, or to provide shelter belts, would be even more divergent. What are the consequences of this for habitat conservation and restoration? We are not sure that protection and planting of corridors are having positive effects on persistence of fauna, but there is prominent ecological theory indicating that this should be the case (Saunders & Hobbs 1991; Soulé & Gilpin 1991). However, this study suggests that linear strips do provide poorer quality habitat than large remnants. Landscape designers must be encouraged to ensure that the intuitive appeal of corridors does not result in the neglect of appropriate management and restoration of large areas of native vegetation. It is also essential that we recognize the difference between planting trees to combat soil erosion and salinity, and planting trees for nature conservation. It will not be constructive to continue to foster the approach that any tree planted for land care benefit will have spin-offs for nature conservation.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

We thank Dominic Sivertsen and Lisa Metcalfe of the NSW National Parks and Wildlife Service for providing a digitized vegetation map. Figure 1 was generated from their data. We are grateful to Paul Wells and Andrew Deane of State Forests for their support, particularly for allowing us to work in the State Forests. All birds were banded under permit from the Australian Bird and Bat Banding Scheme. The manuscript has been improved by the helpful comments of Graham Pyke.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • Andrén, H. & Angelstam, P. (1988) Elevated predation rates as an edge effect in habitat islands: experimental evidence. Ecology, 69, 544547.
  • Barrett, G.W., Ford, H.A. & Recher, H.F. (1994) Conservation of woodland birds in a fragmented rural landscape. Pacific Conservation Biology, 1, 245256.
  • Boles, W. (1988) The Robins and Flycatchers of Australia. Angus and Robertson, North Ryde, Australia.
  • Clarke, K.R. & Warwick, R.M. (1994) Change in Marine Communities: An Approach to Statistical Analysis and Interpretation. Natural Environment Research Council, Plymouth, UK.
  • Donovan, T., Lamberson, R., Kimber, A., Thompson, F. & Faaborg, J. (1995a) Modeling the effects of habitat fragmentation on source and sink demography of neotropical migrant birds. Conservation Biology, 9, 13961407.
  • Donovan, T.M., Thompson, F.R., Faaborg, J. & Probst, J.R. (1995b) Reproductive success of migratory birds in habitat sources and sinks. Conservation Biology, 9, 13801395.
  • Egan, K.H., Farrell, J.R. & Pepper-Edwards, D.L. (1997) Historical and seasonal changes in the community of forest birds at Longneck Lagoon Nature Reserve, Scheyville, New South Wales. Corella, 21, 116.
  • Frith, H.J. (1983) Reader's Digest Complete Book of Australian Birds, 1st edn. Reader's Digest Services Pty Ltd, Surry Hills, Australia.
  • Gardner, J.L. (1998) Experimental evidence for edge-related predation in a fragmented agricultural landscape. Australian Journal of Ecology, 23, 311321.
  • Garnett, S. (1992) Threatened and Extinct Birds of Australia. RAOU report number 82. Royal Australasian Ornithologists Union, Moonee Ponds, Australia.
  • Gates, J.E. & Gysel, L.W. (1978) Avian nest dispersion and fledging success in field–forest ecotones. Ecology, 59, 871883.
  • Gibbs, J.P. (1991) Avian nest predation in tropical wet forest: an experimental study. Oikos, 60, 155161.
  • Greenwood, P. & Harvey, P. (1982) The natal and breeding dispersal of birds. Annual Review of Ecology and Systematics, 13, 121.
  • Hobbs, R.J. (1992) Corridors for conservation: solution or bandwagon? Trends in Ecology and Evolution, 7, 389392.
  • Hoskin, E.S., Hindwood, K.A. & McGill, A.R. (1991) Birds of Sydney, 2nd edn. Surrey Beatty & Sons, Chipping Norton, Australia.
  • Howe, R. & Davis, G. (1991) The demographic significance of ‘sink’ populations. Biological Conservation, 57, 239255.
  • Lynch, J.F. & Saunders, D.A. (1991) Responses of bird species to habitat fragmentation in the wheatbelt of Western Australia: interiors, edges and corridors. Nature Conservation 2: The Role of Corridors (eds D.A.Saunders & R.J.Hobbs), pp. 143158. Surrey Beatty & Sons, Chipping Norton, Australia.
  • Major, R.M., Christie, F.J., Gowing, G. & Ivison, T.J. (1999) Elevated rates of predation on artificial nests in linear strips of habitat. Journal of Field Ornithology, 70, 351364.
  • Møller, A.P. (1989) Nest site selection across field–woodland ecotones: the effect of nest predation. Oikos, 56, 240246.
  • Murcia, C. (1995) Edge effects in fragmented forests: implications for conservation. Trends in Ecology and Evolution, 10, 5862.
  • Pulliam, H. & Danielson, B. (1991) Sources, sinks and habitat selection: a landscape perspective on population dynamics. American Naturalist, 137, 6066.
  • Ratti, J.T. & Reese, K.P. (1988) Preliminary test of the ecological trap hypothesis. Journal of Wildlife Management, 52, 484491.
  • Recher, H.F. & Lim, L. (1990) A review of current ideas of the extinction, conservation and management of Australia's terrestrial vertebrate fauna. Proceedings of the Ecological Society of Australia, 16, 287301.
  • Robinson, D. (1991) Threatened birds in Victoria: their distributions, ecology and future. Victorian Naturalist, 108, 6777.
  • Robinson, D. (1993) Vale Toolern Vale: the loss of our woodland birds. Wingspan, 9, 20.
  • Saunders, D.A. (1989) Changes in the avifauna of a region, district and remnant as a result of fragmentation of native vegetation: the wheatbelt of Western Australia. A case study. Biological Conservation, 50, 90135.
  • Saunders, D.A. & Hobbs, R.J. (1991) The role of corridors in conservation: what do we know and where do we go? Nature Conservation 2: The Role of Corridors (eds D.A.Saunders & R.J.Hobbs), pp. 421427. Surrey Beatty & Sons, Chipping Norton, Australia.
  • Saunders, D.A. & Ingram, J.A. (1995) Birds of Southwestern Australia: An Atlas of Changes in the Distribution and Abundance of the Wheatbelt Avifauna. Surrey Beatty & Sons, Chipping Norton, Australia.
  • Saunders, D.A., Hobbs, R.J. & Ehrlich, P.R. (1993) Nature Conservation 3: Reconstruction of Fragmented Ecosystems: Problems and Possibilities. Surrey Beatty & Sons, Chipping Norton, Australia.
  • Sherry, T. & Holmes, R. (1989) Age-specific social dominance affects habitat use by breeding American redstarts (Setophaga ruticilla): a removal experiment. Behavioral Ecology and Sociobiology, 25, 327333.
  • Simberloff, D., Farr, J.A., Cox, J. & Mehlman, D.W. (1992) Movement corridors: conservation bargains or poor investments? Conservation Biology, 6, 493504.
  • Sisk, T.D. & Margules, C.R. (1993) Habitat edges and restoration: methods for quantifying edge effects and predicting the results of restoration efforts. Nature Conservation 3: Reconstruction of Fragmented Ecosystems (eds D.A.Saunders, R.J.Hobbs & P.R.Ehrlich), pp. 5769. Surrey Beatty & Sons, Chipping Norton, Australia.
  • Sivertsen, D. & Metcalfe, L. (1995) Natural vegetation of the southern wheat-belt (Forbes and Cargelligo 1 : 250 000 map sheets). Cunninghamia, 4, 103128.
  • Soulé, M.E. & Gilpin, M.E. (1991) The theory of wildlife corridor capability. Nature Conservation 2: The Role of Corridors (eds D.A.Saunders & R.J.Hobbs), pp. 38. Surrey Beatty & Sons, Chipping Norton, Australia.
  • Temple, S.A. & Carey, J.R. (1988) Modeling dynamics of habitat–interior bird populations in fragmented landscapes. Conservation Biology, 2, 340347.
  • United Nations Environment Programme (1995) Global Biodiversity Assessment. Cambridge University Press, Cambridge, UK.
  • Villard, M., Martin, P. & Drummond, C. (1993) Habitat fragmentation and pairing success in the ovenbird (Seiurus aurocapillus). Auk, 110, 759768.

Received 13 October 1998; revision received 24 June 1999